Harden backup and restore reliability paths

liamvibecodes · liamvibecodes · commit f52cb4555c93 · 2026-02-22T01:18:58.000-05:00
diff --git a/README.md b/README.md
@@ -29,7 +29,7 @@ This tool backs up configs, databases, and compose files every night. When somet
 | Item | Details |
 |------|---------|
 | **Config files** | config.xml, config.yml, settings.json, *.conf from each service |
-| **Databases** | All *.db files (Sonarr, Radarr, Prowlarr, Lidarr, Bazarr, etc.) |
+| **Databases** | All *.db files (Sonarr, Radarr, Prowlarr, Lidarr, Bazarr, etc.), snapshot-backed with sqlite3 when available |
 | **Compose file** | docker-compose.yml |
 | **Environment** | .env (sensitive values redacted) |
 | **Container state** | Snapshot of running containers at backup time |
@@ -85,6 +85,7 @@ bash restore.sh 20260222-020000
 ```
 
 The restore process stops your containers, copies configs and databases back, restarts everything, and runs a health check.
+If container restart fails, the script exits non-zero and prints clear remediation steps instead of claiming success.
 
 **Note:** .env values are redacted in backups for security. After restoring, check your .env file and re-add any secrets if needed.
 
diff --git a/backup.sh b/backup.sh
@@ -21,13 +21,27 @@ usage() {
     echo "  --path DIR     Media directory (default: ~/Media)"
     echo "  --keep DAYS    Days of backups to keep (default: 14)"
     echo "  --help         Show this help"
-    exit 0
+    exit "${1:-0}"
 }
 
 while [[ $# -gt 0 ]]; do
     case "$1" in
-        --path) MEDIA_DIR="$2"; shift 2 ;;
-        --keep) KEEP_DAYS="$2"; shift 2 ;;
+        --path)
+            if [[ $# -lt 2 ]]; then
+                echo -e "${RED}ERR${NC} Missing value for --path"
+                usage 1
+            fi
+            MEDIA_DIR="$2"
+            shift 2
+            ;;
+        --keep)
+            if [[ $# -lt 2 ]]; then
+                echo -e "${RED}ERR${NC} Missing value for --keep"
+                usage 1
+            fi
+            KEEP_DAYS="$2"
+            shift 2
+            ;;
         --help) usage ;;
         *) echo -e "${RED}ERR${NC} Unknown option: $1"; exit 1 ;;
     esac
@@ -61,7 +75,12 @@ if [[ ! -d "$MEDIA_DIR" ]]; then
     exit 1
 fi
 
-mkdir -p "$BACKUP_DIR" "$LOG_DIR" "$BACKUP_STAGING"
+if ! [[ "$KEEP_DAYS" =~ ^[0-9]+$ ]]; then
+    echo -e "${RED}ERR${NC}  --keep must be a whole number"
+    exit 1
+fi
+
+mkdir -p "$BACKUP_DIR" "$LOG_DIR" "$BACKUP_STAGING" "$BACKUP_STAGING/configs" "$BACKUP_STAGING/databases"
 
 # ==============================
 # Config files
@@ -91,13 +110,30 @@ echo ""
 # ==============================
 echo -e "${CYAN}--- Databases ---${NC}"
 
+# Use sqlite online backup when available to avoid inconsistent hot copies.
+backup_db_file() {
+    local src="$1"
+    local dst="$2"
+    local rel="$3"
+
+    if command -v sqlite3 &>/dev/null; then
+        if sqlite3 "$src" ".timeout 5000" ".backup \"$dst\"" >/dev/null 2>&1; then
+            echo -e "${GREEN}OK${NC}   Snapshot database: $rel"
+            return 0
+        fi
+        echo -e "${YELLOW}WRN${NC}  sqlite3 snapshot failed for $rel, falling back to file copy"
+    fi
+
+    cp "$src" "$dst"
+    echo -e "${GREEN}OK${NC}   Copied database: $rel"
+}
+
 find "$MEDIA_DIR" -maxdepth 3 -type f -name "*.db" \
     ! -path "*/backups/*" ! -path "*/logs/*" 2>/dev/null | while read -r file; do
     rel_path="${file#$MEDIA_DIR/}"
     dest_dir="$BACKUP_STAGING/databases/$(dirname "$rel_path")"
     mkdir -p "$dest_dir"
-    cp "$file" "$dest_dir/"
-    echo -e "${GREEN}OK${NC}   Copied database: $rel_path"
+    backup_db_file "$file" "$dest_dir/$(basename "$file")" "$rel_path"
 done
 
 DB_COUNT=$(find "$BACKUP_STAGING/databases" -type f 2>/dev/null | wc -l | tr -d ' ')
@@ -166,11 +202,9 @@ echo ""
 # ==============================
 echo -e "${CYAN}--- Pruning Old Backups ---${NC}"
 
-PRUNED=0
 find "$BACKUP_DIR" -name "backup-*.tar.gz" -type f -mtime +"$KEEP_DAYS" 2>/dev/null | while read -r old; do
     rm -f "$old"
     echo -e "${YELLOW}DEL${NC}  Removed $(basename "$old")"
-    PRUNED=$((PRUNED + 1))
 done
 
 REMAINING=$(find "$BACKUP_DIR" -name "backup-*.tar.gz" -type f 2>/dev/null | wc -l | tr -d ' ')
diff --git a/install.sh b/install.sh
@@ -33,12 +33,19 @@ usage() {
     echo "  bash install.sh              # Install nightly backup at 2am"
     echo "  bash install.sh --hour 3     # Install nightly backup at 3am"
     echo "  bash install.sh --uninstall  # Remove scheduled backup"
-    exit 0
+    exit "${1:-0}"
 }
 
 while [[ $# -gt 0 ]]; do
     case "$1" in
-        --hour) HOUR="$2"; shift 2 ;;
+        --hour)
+            if [[ $# -lt 2 ]]; then
+                echo -e "${RED}ERR${NC} Missing value for --hour"
+                usage 1
+            fi
+            HOUR="$2"
+            shift 2
+            ;;
         --uninstall) UNINSTALL=true; shift ;;
         --help) usage ;;
         *) echo -e "${RED}ERR${NC} Unknown option: $1"; exit 1 ;;
@@ -84,6 +91,11 @@ if [[ ! -f "$BACKUP_SCRIPT" ]]; then
     exit 1
 fi
 
+if ! [[ "$HOUR" =~ ^[0-9]+$ ]]; then
+    echo -e "${RED}ERR${NC}  Hour must be a whole number between 0 and 23"
+    exit 1
+fi
+
 if [[ "$HOUR" -lt 0 || "$HOUR" -gt 23 ]]; then
     echo -e "${RED}ERR${NC}  Hour must be between 0 and 23"
     exit 1
diff --git a/restore.sh b/restore.sh
@@ -30,14 +30,21 @@ usage() {
     echo "  bash restore.sh --latest"
     echo "  bash restore.sh 20260222-020000"
     echo "  bash restore.sh backup-20260222-020000.tar.gz"
-    exit 0
+    exit "${1:-0}"
 }
 
 while [[ $# -gt 0 ]]; do
     case "$1" in
         --list) LIST_MODE=true; shift ;;
         --latest) LATEST_MODE=true; shift ;;
-        --path) MEDIA_DIR="$2"; shift 2 ;;
+        --path)
+            if [[ $# -lt 2 ]]; then
+                echo -e "${RED}ERR${NC} Missing value for --path"
+                usage 1
+            fi
+            MEDIA_DIR="$2"
+            shift 2
+            ;;
         --help) usage ;;
         -*) echo -e "${RED}ERR${NC} Unknown option: $1"; exit 1 ;;
         *) BACKUP_TARGET="$1"; shift ;;
@@ -63,15 +70,15 @@ if $LIST_MODE; then
     echo ""
 
     FOUND=0
-    for f in $(ls -t "$BACKUP_DIR"/backup-*.tar.gz 2>/dev/null); do
+    while IFS= read -r f; do
         name=$(basename "$f")
         size=$(du -sh "$f" | cut -f1)
         date_part=$(echo "$name" | sed 's/backup-\([0-9]*\)-\([0-9]*\).*/\1/')
         time_part=$(echo "$name" | sed 's/backup-[0-9]*-\([0-9]*\).*/\1/')
         formatted_date="${date_part:0:4}-${date_part:4:2}-${date_part:6:2} ${time_part:0:2}:${time_part:2:2}:${time_part:4:2}"
         echo -e "  ${GREEN}$name${NC}  ($size)  $formatted_date"
         FOUND=$((FOUND + 1))
-    done
+    done < <(ls -1t "$BACKUP_DIR"/backup-*.tar.gz 2>/dev/null)
 
     if [[ $FOUND -eq 0 ]]; then
         echo -e "  ${YELLOW}No backups found${NC}"
@@ -142,8 +149,12 @@ trap cleanup EXIT
 # ==============================
 echo -e "${CYAN}--- Extracting ---${NC}"
 tar -xzf "$BACKUP_FILE" -C "$TEMP_DIR"
-BACKUP_ROOT=$(ls "$TEMP_DIR")
-EXTRACTED="$TEMP_DIR/$BACKUP_ROOT"
+BACKUP_ROOT=$(find "$TEMP_DIR" -mindepth 1 -maxdepth 1 -type d | head -1)
+if [[ -z "$BACKUP_ROOT" || ! -d "$BACKUP_ROOT" ]]; then
+    echo -e "${RED}ERR${NC}  Backup archive did not contain a valid root directory"
+    exit 1
+fi
+EXTRACTED="$BACKUP_ROOT"
 echo -e "${GREEN}OK${NC}   Extracted to temp directory"
 echo ""
 
@@ -152,9 +163,19 @@ echo ""
 # ==============================
 echo -e "${CYAN}--- Stopping Containers ---${NC}"
 
-if command -v docker &>/dev/null && [[ -f "$MEDIA_DIR/docker-compose.yml" || -f "$MEDIA_DIR/docker-compose.yaml" ]]; then
-    (cd "$MEDIA_DIR" && docker compose stop 2>/dev/null) || true
-    echo -e "${GREEN}OK${NC}   Containers stopped"
+STOP_COMPOSE_FILE=""
+if [[ -f "$MEDIA_DIR/docker-compose.yml" ]]; then
+    STOP_COMPOSE_FILE="$MEDIA_DIR/docker-compose.yml"
+elif [[ -f "$MEDIA_DIR/docker-compose.yaml" ]]; then
+    STOP_COMPOSE_FILE="$MEDIA_DIR/docker-compose.yaml"
+fi
+
+if command -v docker &>/dev/null && [[ -n "$STOP_COMPOSE_FILE" ]]; then
+    if (cd "$MEDIA_DIR" && docker compose stop >/dev/null 2>&1); then
+        echo -e "${GREEN}OK${NC}   Containers stopped"
+    else
+        echo -e "${YELLOW}WRN${NC}  Failed to stop containers cleanly, continuing restore"
+    fi
 else
     echo -e "${YELLOW}WRN${NC}  No compose file or Docker not found, skipping container stop"
 fi
@@ -239,15 +260,27 @@ echo ""
 # ==============================
 echo -e "${CYAN}--- Restarting Containers ---${NC}"
 
-if command -v docker &>/dev/null && [[ -f "$MEDIA_DIR/docker-compose.yml" || -f "$MEDIA_DIR/docker-compose.yaml" ]]; then
-    (cd "$MEDIA_DIR" && docker compose up -d 2>/dev/null) || true
-    echo -e "${GREEN}OK${NC}   Containers started"
-    echo ""
+START_COMPOSE_FILE=""
+if [[ -f "$MEDIA_DIR/docker-compose.yml" ]]; then
+    START_COMPOSE_FILE="$MEDIA_DIR/docker-compose.yml"
+elif [[ -f "$MEDIA_DIR/docker-compose.yaml" ]]; then
+    START_COMPOSE_FILE="$MEDIA_DIR/docker-compose.yaml"
+fi
+
+if command -v docker &>/dev/null && [[ -n "$START_COMPOSE_FILE" ]]; then
+    if (cd "$MEDIA_DIR" && docker compose up -d >/dev/null 2>&1); then
+        echo -e "${GREEN}OK${NC}   Containers started"
+        echo ""
 
-    # Health check
-    echo -e "${CYAN}--- Health Check ---${NC}"
-    sleep 3
-    (cd "$MEDIA_DIR" && docker compose ps 2>/dev/null) || true
+        # Health check
+        echo -e "${CYAN}--- Health Check ---${NC}"
+        sleep 3
+        (cd "$MEDIA_DIR" && docker compose ps 2>/dev/null) || true
+    else
+        echo -e "${RED}ERR${NC}  Restore completed, but container restart failed."
+        echo -e "${RED}ERR${NC}  Check compose syntax/secrets and run: (cd \"$MEDIA_DIR\" && docker compose up -d)"
+        exit 1
+    fi
 else
     echo -e "${YELLOW}WRN${NC}  No compose file or Docker not found, skipping restart"
 fi
